Given the high personal and economic costs of stroke, significant resources have been devoted to rehabilitation efforts within the VA and elsewhere. Despite this emphasis, all too often recovery after stroke remains partial, and many patients do not respond to traditional therapies. The difficulties in improving recovery after stroke stem in part from the fact that traditional approaches to predicting the effects of stroke-related brain lesions on subsequent function are often imprecise, particularly so for higher cognitive functions. Recent advances have demonstrated, for example, that similar-appearing lesions may give rise to disparate phenotypes based upon the extent to which they disrupt specific large-scale brain networks. Thus, in this proposal we will take advantage of advances in MRI methodology and analytics, within the context of validated behavioral metrics and new statistical techniques, to develop new predictors of functional recovery after stroke. Over the course of the study, patients referred from acute care hospitals to the CREC in Martinez, California for rehabilitation after stroke will be recruited to participate within two weeks of their index event. Those who provide informed consent will undergo a battery of tests to assess cognitive, emotional, motor, and other neurological function. In parallel, they will undergo structural MRI, resting state functional MRI (rs-fMRI) and diffusion tractography imaging (DTI) from which connectivity metrics derived through graph theory and Granger causality will be determined. Both behavioral and neuroimaging data will be obtained at three time points: within two weeks of the sentinel event, at three months, and at twelve months. Following the acquisition of these behavioral and imaging metrics, advanced statistical methods will be used to search for validated predictors of cognitive, emotional, and other neurological recovery at three and twelve months after stroke. As such, this proposal takes advantage of (1) validated behavioral and cognitive measures; (2) a new connectivity brain science that permits the quantification of the integrity of brain networks and has given rise to hypotheses about their evolution after injury; (3) advanced statistical techniques; and (4) longitudinal assessments in order to identify markers that will help to predict recovery after stroke. This work hopefully represents the first step in a long-term program designed to address the significant personal and economic costs of stroke in veterans and others. In addition to permitting prospective validation of any predictors of cognitive recovery, these results may also form the basis for the assessment of future approaches to stroke treatment, including individualized medication trials and targeted non-invasive brain stimulation to enhance rehabilitation efforts.